Cellulose-Derived Oligomers Act as Damage-Associated Molecular Patterns and Trigger Defense-Like Responses

Cellobiose, a danger signal derived from breakdown of the major cell wall polymer cellulose, enhances plant defenses triggered by microbe-derived elicitors.

Abstract

The plant cell wall, often the site of initial encounters between plants and their microbial pathogens, is composed of a complex mixture of cellulose, hemicellulose, and pectin polysaccharides as well as proteins. The concept of damage-associated molecular patterns (DAMPs) was proposed to describe plant elicitors like oligogalacturonides (OGs), which can be derived by the breakdown of the pectin homogalacturon by pectinases. OGs act via many of the same signaling steps as pathogen- or microbe-associated molecular patterns (PAMPs) to elicit defenses and provide protection against pathogens. Given both the complexity of the plant cell wall and the fact that many pathogens secrete a wide range of cell wall-degrading enzymes, we reasoned that the breakdown products of other cell wall polymers may be similarly biologically active as elicitors and may help to reinforce the perception of danger by plant cells. Our results indicate that oligomers derived from cellulose are perceived as signal molecules in Arabidopsis (Arabidopsis thaliana), triggering a signaling cascade that shares some similarities to responses to well-known elicitors such as chitooligomers and OGs. However, in contrast to other known PAMPs/DAMPs, cellobiose stimulates neither detectable reactive oxygen species production nor callose deposition. Confirming our idea that both PAMPs and DAMPs are likely to cooccur at infection sites, cotreatments of cellobiose with flg22 or chitooligomers led to synergistic increases in gene expression. Thus, the perception of cellulose-derived oligomers may participate in cell wall integrity surveillance and represents an additional layer of signaling following plant cell wall breakdown during cell wall remodeling or pathogen attack.

The primary plant cell wall is composed of a complex interconnected mixture of proteins and polysaccharides, mainly cellulose, hemicellulose, and pectin. The secondary cell wall also contains lignin. These strong polymeric networks provide structural integrity to plant cells and protection from the external environment (Somerville et al., 2004). To gain access to the cell cytoplasm, pathogens have to overcome the plant cell wall barrier and do so mainly through the secretion of cell wall-degrading enzymes (Howard, 1997; Toth and Birch, 2005). Plants can perceive the presence of pathogens at the cell surface via the recognition of conserved microbial molecules, named pathogen- or microbe-associated molecular patterns (PAMPs). Well-studied examples of PAMPs are the elicitor-active peptides of bacterial flagellin (flg22), the bacterial elongation factor EF-Tu (elf18), and chitooligomers, breakdown products of fungal cell walls and insect exoskeletons (Kunze et al., 2004; Chinchilla et al., 2006; Miya et al., 2007; Boller and Felix, 2009). The recognition of such molecules is achieved by specific plasma membrane-resident receptors called pattern recognition receptors (PRRs). Upon PAMP perception, a signaling cascade is initiated to activate plant defense responses in a process termed pattern-triggered immunity (PTI; Jones and Dangl, 2006). PTI is characterized by the influx of calcium ions, the generation of reactive oxygen species (ROS; Alonso et al., 2003), the activation of mitogen-activated protein kinases (MAPKs; Torres et al., 2002; Pitzschke et al., 2009; Tena et al., 2011), and changes in gene expression leading to the increased production of defense compounds and proteins, thus equipping the plant cell to defend itself.

While gaining access to the cytoplasm of plant cells during the penetration phase, pathogens breach the cell walls, releasing host peptides and oligosaccharide fragments (i.e. damage-associated molecular patterns [DAMPs]; Howard, 1997). Pectin-derived oligogalacturonides (OGs) are well-characterized DAMPs capable of activating plant immunity (Kohorn et al., 2009; Brutus et al., 2010). OGs are perceived by WAK1 and WAK2, cell wall-associated receptor-like kinases required for cell expansion (Kohorn et al., 2006). Transgenic plants overexpressing WAK1 are more resistant to the necrotrophic fungal pathogen Botrytis cinerea (Brutus et al., 2010).

Plants also encode a wide array of cell wall-degrading enzymes, which are thought to play a role in cell wall remodeling during growth and development (Cosgrove, 2005). Given this dynamic and complex nature of the plant cell wall and the diversity of cell wall-degrading/modifying enzymes encoded by many pathogens, there are a multitude of small molecules that may be generated at the infection court. Such small molecules have the potential to be recognized as danger signals and to be perceived by a cell wall integrity-sensing system (Pilling and Höfte, 2003; Vorwerk et al., 2004; Hématy et al., 2009; Bolouri Moghaddam and Van den Ende, 2012; Wolf et al., 2012). Experimental evidence has accumulated over the past decade to support the idea that plants monitor the status of the cell wall via a cell wall integrity-sensing system (Hématy et al., 2007; Cheung and Wu, 2011; Denness et al., 2011; Ramírez et al., 2011). Despite the progress in the field, our understanding of the cell wall-derived signals and molecular mechanisms underlying the recognition of cell wall damage is limited. Cell walls are anchored to the cell surface via the cell wall biosynthetic machinery and by structural and sensory proteins that bind to cell wall components and maintain plasma membrane-cell wall contacts (Liu et al., 2015). This link is thought to be essential for plant development and responses to external stimuli (Wolf et al., 2012). Cellulose is synthesized at the plasma membrane by the cellulose synthase complex, which converts UDP-Glc into β-1,4-glucan chains that crystallize into cellulose microfibrils in the cell wall. Cellulose microfibrils are the major load-bearing components of the plant cell wall. Thus, loss of cellulose microfibril integrity has drastic effects on plant cells (Somerville, 2006).

Here, we present work demonstrating that the perception of cellulose degradation products, in the absence of catastrophic cell wall damage and the loss of cellular integrity observed in previous studies, activates defense responses similar to PTI in Arabidopsis (Arabidopsis thaliana). Furthermore, cotreatment of cellulose fragments and PAMPs like flg22 or chitooligomers leads to synergistic increases in gene expression, suggesting that plant cells may be able to respond defensively earlier and at lower doses of mixtures of elicitors likely to be found in the infection court.

RESULTS

Defense-Related WRKY Transcription Factors Are Up-Regulated by Cellulose Oligomer Treatment

WRKY transcription factors have long been implicated in the regulation of plant responses to biotic and abiotic stresses, and frequently, a single WRKY transcription factor regulates the transcriptional reprogramming of multiple plant processes (Rushton et al., 2010). Using publicly available gene expression data sets, we selected several WRKY transcription factors with elevated transcript levels after chitooligomer treatment, suggesting an active role in the defense response, for further study. Transgenic seedlings expressing WRKY promoter:GUS constructs (WRKYp:GUS) were used to identify cell wall-derived oligosaccharides that were capable of stimulating higher expression of the selected defense-related WRKY genes. We found that oligomers of cellulose (DP2 and DP3) caused enhanced expression of the GUS gene under the control of WRKY30 and WRKY40 promoters (Fig. 1). We determined the time course of expression of WRKY30, WRKY40, and other defense-related WRKY genes by qRT-PCR, and the results showed that WRKY30 had the strongest transcriptional response of all WRKY genes tested, peaking at 25 min after treatment with cellobiose (Fig. 1). Treatment with cellobiose (DP2), cellotriose (DP3), and cellotetraose (DP4) elicited similar levels of WRKY30 expression (Supplemental Fig. S1). This observation, along with reports that two classes of cellulases (i.e. GH6 and GH7) commonly found in saprophytic and hemibiotrophic fungi produce cellobiose (Spanu et al., 2010; Glass et al., 2013), prompted us to continue using cellobiose as a representative cellulose degradation product. Cellobiose treatment triggered the enhanced expression of WRKY30 in seedling roots and seedling shoots; however, WRKY30 also exhibited constitutive expression in cotyledons (Fig. 1). The regulation of WRKY30 expression in seedling roots was tightly regulated, being elicitor dependent and undetectable in the absence of a stimulus (Supplemental Fig. S2). Therefore, WRKY30 expression in seedling roots at 25 min posttreatment was used as a molecular marker for the further characterization of plant responses to cellobiose.

Figure 1.

Expression patterns of defense-related WRKY transcription factors after elicitor treatment. A, Representative GUS expression patterns in the primary root of transgenic, 7-d-old Arabidopsis seedlings harboring WRKY30p:GUS and WRKY40p:GUS fusions. Elicitors are indicated. B, Quantitative reverse transcription (qRT)-PCR results of wild-type 7-d-old Arabidopsis whole seedlings treated with 100 μm cellobiose harvested at different times after treatment. Expression values are relative to untreated controls. Error bars represent sd of two biological replicates with three technical replicates each. The experiment was repeated twice with similar results.

WRKY30 Is Induced by β-1,4-Glucan Oligosaccharides

Soluble sugars such as Suc, raffinose, and trehalose can play a signaling role in plant innate immunity (Bolouri Moghaddam and Van den Ende, 2012). For example, Suc treatment leads to the induction of pathogenesis-related genes (Solfanelli et al., 2006) and to strong enhanced expression of genes in the anthocyanin biosynthetic pathway (Solfanelli et al., 2006). Synthesis of the nonreducing Glc disaccharide trehalose (α-1,1-diglucose) has been shown to regulate responses to environmental stresses (Iordachescu and Imai, 2008). In addition, trehalose synthesis by Pseudomonas aeruginosa strain PA14, a multihost pathogen that infects plants, nematodes, insects, and vertebrates, is required for full virulence on Arabidopsis (Djonović et al., 2013). Therefore, it is possible that plants have evolved to recognize apoplastic trehalose as a defense mechanism. Given that trehalose is also a Glc disaccharide, like cellobiose (β-1,4-diglucose), we asked whether the responses to cellobiose were unique to this Glc dimer or not. We exposed WRKY30p:GUS transgenic seedlings to a panel of disaccharides with various linkages and found that, in seedling roots, WRKY30p:GUS expression was elicited exclusively by cellobiose among all the sugars tested (Fig. 2). Glc did not induce WRKY30p:GUS expression in these tests, indicating that the observed cellobiose responses are not due to the breakdown of cellobiose to Glc. These results suggest that a specific receptor for small oligomers of cellulose may exist in Arabidopsis seedlings.

Figure 2.

GUS expression patterns in wild-type 7-d-old Arabidopsis seedlings harboring the WRKY30p:GUS construct in response to Glc and various disaccharide treatments. All treatments were applied at 100 μm for 16 h. A minimum of 16 seedlings were tested in each treatment. Representative seedlings are shown.

WRKY30 has been characterized as a general stress-responsive gene (Scarpeci et al., 2013), and the work presented here shows that its expression is stimulated in seedlings by several PAMP/DAMP elicitors including chitooligomers, which are oligomers of β-1,4-N-acetyl-d-glucosamine (Supplemental Figs. S2 and S3). The LysM receptor-like kinase (CERK1) binds to chitooligomers and is required for chitin and peptidoglycan perception in Arabidopsis (Miya et al., 2007; Wan et al., 2008; Willmann et al., 2011), and plants carrying a mutation in CERK1 can no longer respond to chitin stimulation. The cerk1 null mutant still responds to cellobiose treatment, indicating that cellobiose does not promiscuously activate this receptor (Supplemental Fig. S3).

Cellobiose Induces Responses Elicited by Other PAMPs/DAMPs

Cellobiose Treatment Triggers an Early Calcium Transient

Calcium is a ubiquitous and protean intracellular second messenger. A wide range of stimuli cause changes in intracellular calcium concentration in plants (Sanders et al., 1999). These changes generate unique stimulus-dependent calcium signatures (i.e. timing and magnitude of signal) leading to multiple physiological responses (Sanders et al., 1999, 2002; Lecourieux et al., 2005). Intracellular calcium transients have been shown to occur after exposure to pathogens or purified elicitors and, therefore, are considered one of the hallmarks of PAMP/DAMP perception (Allen et al., 2001; Ma et al., 2012, 2013; Michal Johnson et al., 2014). We used aequorin-expressing Arabidopsis seedlings (Knight et al., 1991) to determine if cellobiose exposure generated a calcium response (Fig. 3). Our results show that cellobiose exposure generates a fast and short-lived intracellular calcium elevation, lasting for only about 200 s, with levels peaking at 100 s posttreatment (Fig. 3). Plants pretreated with the calcium chelator EGTA (2.5 mm) showed a 60% reduction in WRKY30 expression after cellobiose treatment, indicating that the calcium transient is part of the cellobiose-generated signaling cascade leading to the activation of gene expression (Fig. 3). Control treatments using Glc and Suc did not elicit a calcium response (Supplemental Fig. S4), highlighting the specificity of this response to cellobiose.

Figure 3.

Cellobiose-generated intracellular calcium influx. A, Aequorin-expressing plants were treated with cellobiose and immediately visualized using a CCD camera. To control for aequorin presence, at the end of the experiment, remaining aequorin was discharged by the addition of an equal volume of solution containing 2 m CaCl₂ and 20% (v/v) ethanol. B, Pixel intensities of images similar to those in A were quantified using ImageJ. Mean and se are shown (n = 18). C, qRT-PCR results showing WRKY30 expression in 7-d-old seedling roots pretreated with 2.5 mm EGTA in response to cellobiose treatment. Means and sd of three biological replicates are shown. Asterisk indicates significant difference from WRKY30 expression observed in cellobiose-treated Columbia-0 (Col-0) plants (P = 0.05 by one-way ANOVA coupled to Tukey's test).

Cellobiose Treatment Activates MAPKs

MAPK cascades are central to innate immune signaling (Asai et al., 2002; Meng and Zhang, 2013). MPK3, MPK6, MPK4, and MPK11 are strongly activated upon PAMP/DAMP treatment (Meng and Zhang, 2013). We tested whether cellobiose treatment also would lead to MAPK activation. Cellobiose treatment activates MPK6 and MPK3 at very early time points, with stronger activation of MPK6 (Fig. 4; Supplemental Fig. S5). Phosphorylation of MPK6 was visible between 5 and 15 min, being strongest at 10 min after induction (Fig. 4). Elevated expression of WRKY30 by cellobiose is decreased 20-fold in the mpk6-2 mutant (Salk_073907), indicating that MPK6 plays an important role in the cellobiose signal transduction pathway leading to WRKY30 expression (Fig. 4).

Figure 4.

MAPK activation by cellobiose treatments. A, Western blot showing the early activation of MPKs after cellobiose treatment; MPK6 is the most strongly activated. B, WRKY30 expression after 100 μm cellobiose treatment was assessed in the mpk6-2 mutant background. Error bars represent sd of three biological replicates. Asterisks represent points that differ significantly from WRKY30 expression observed in cellobiose-treated Col-0 plants (P < 0.05 by one-way ANOVA coupled to Tukey’s test).

Global Arabidopsis Gene Expression Profiles Are Similar after Cellobiose, Chitooligomer, or OG Treatment

Studies of the global suite of differentially regulated genes after PAMP/DAMP elicitor treatment have highlighted the high degree of overlap in the early transcriptional response following elicitor perception, indicating that a basal broad-spectrum response is a common feature following the recognition of a danger signal (Zhang et al., 2002; Moscatiello et al., 2006; Zipfel et al., 2006; Denoux et al., 2008; Wan et al., 2008). However, it also has been shown that the profile of early signaling events, including the kinetics of transcriptional changes following elicitor treatment, varies between elicitors (Garcia-Brugger et al., 2006; Zipfel et al., 2006; Denoux et al., 2008). With this in mind, we performed an Affymetrix microarray experiment on Arabidopsis seedling roots treated with chitooligomers, OGs, or cellobiose for 25-min (early) and 3-h (late) time points. Chitooligomers and OGs were used in saturating concentrations (Hu et al., 2004; Miya et al., 2007; Shinya et al., 2012). No group of genes was substantially up-regulated exclusively by cellobiose and OGs (i.e. DAMPs) but not by chitooligomers (i.e. PAMPs), a characteristic we would expect for genes encoding cell wall integrity sensing and response. Instead, our results indicate that the early transcriptional response to cellobiose is similar to that following treatment with other known PAMP/DAMPs. At the 25-min time point, cellobiose-triggered changes overlapped more strongly with those elicited by the pathogen-derived chitooligomers than by plant-derived OGs (Fig. 5; Supplemental Table S1). In the group of genes up-regulated more than 2.5-fold after 25 min of treatment, chitooligomer treatment elicited the largest number of transcriptional changes (735 genes), followed by cellobiose (689) and OGs (568), with 506 genes similarly induced by all three treatments. Nonetheless, as stated above, cellobiose elicitation of the marker gene WRKY30 expression is independent of the chitin receptor CERK1; thus, the similarity observed between cellobiose and chitooligomer-induced transcriptional changes is not due to promiscuous receptor binding. In addition, hierarchical clustering showed higher dissimilarity among gene expression profiles elicited by the three elicitors at 3 h posttreatment and grouped chitooligomer-elicited profiles more closely to OG-elicited profiles; cellobiose was the most dissimilar among the three (Fig. 5; Supplemental Table S2; Nekrasov et al., 2009).

Figure 5.

Results of microarray analysis revealed a high degree of overlap among genes induced following treatment with cellobiose, chitooligomers, or OGs. A, Venn diagrams show that cellobiose samples (CB) exhibited higher overlap with chitin at 25 min. After 3 h, chitin-treated samples had ∼5 times the number of genes with greater than 2.5-fold higher expression levels relative to the other treatments (P < 0.01). B, Hierarchical clustering analysis of global transcriptional changes showed substantial similarity among all three treatments at 25 min, with increasing dissimilarities at 3 h. Numbers inside nodes represent correlation values. The color bar represents fold change values (log₂). CH, Chitin.

Past studies showed that PAMPs can induce the expression of defense genes independently of the defense-associated hormones salicylic acid, jasmonic acid, and ethylene (Zhang et al., 2002; Ferrari et al., 2003, 2007; Zipfel et al., 2004). Conversely, PAMP/DAMP treatment also has been reported to stimulate jasmonic acid and ethylene production (Doares et al., 1995; Simpson et al., 1998) as well as the elevated expression of genes encoding proteins linked to salicylic acid-mediated responses (Denoux et al., 2008). The results from cellobiose-treated seedling roots showed that cellobiose exposure also triggers the up-regulation of genes linked to biosynthesis and signaling mediated by defense hormones after 25 min (Table I; Supplemental Table S1). For example, SAG101 and PAD4, genes associated with salicylic acid signaling, were up-regulated by cellobiose after 25 min. ACS7, an ACC synthase involved in the synthesis of ethylene (Yamagami et al., 2003), was induced 15-fold in the cellobiose-treated samples, similar to the up-regulation found in the chitin and OG samples (17- and 10-fold increases, respectively). LOX3 and LOX4, genes encoding proteins in the octadecanoid pathway leading to the production of jasmonic acid, were up-regulated at 25 min, returning to basal levels after 3 h in all treatments. The magnitude of amplification of the LOX genes at 25 min was more similar between cellobiose and chitin, approximately 20-fold for LOX3 and 100-fold for LOX4 in both treatments, whereas in the OG-treated samples, LOX3 and LOX4 were up 8- and 30-fold, respectively. Genes involved in the biosynthesis of defense-associated indole glucosinolates, including the transcriptional regulator MYB51, were up-regulated by all treatments at 25 min. However, in contrast to what we observed for the cellobiose-treated samples, in which the expression of the indole glucosinolate biosynthetic genes IGMT1/2/3/4 returned to basal levels, the expression of these genes remained up-regulated after 3 h of chitooligomer treatment and, to a lesser extent, also remained up-regulated in the OG-treated samples (Table I). These data are in agreement with previous studies; when comparing the effects of different elicitors on gene expression, the results were more quantitative than qualitative (Denoux et al., 2008). In fact, it appears that the global response to cellobiose and OGs diminishes more rapidly than after chitooligomer treatment; however, we observed a significant reduction in the number of differentially regulated genes at 3 h for all three treatments, highlighting the transient nature of the early broad-spectrum basal defense response.

Table I. Elicitor-induced fold changes of selected genes involved in hormone signaling/biosynthesis and defense-associated processes.

CB, Cellobiose; CH, chitin.

Process	Gene Identifier	Probe	Annotation	Fold Change (P < 0.05)
				25 min after Treatment			3 h after Treatment
				CB	CH	OG	CB	CH	OG
Salicylic acid signaling	At1g74710	262177_at	EDS16/SID2	2.00	1.51	4.79	0.76	0.86	3.49
	At3g48090	252373_at	EDS1	2.13	2.47	1.92	0.90	1.38	1.02
	At3g52430	252060_at	PAD4	6.74	8.59	6.18	1.07	1.58	1.06
	At5g14930	246600_at	SAG101	2.99	4.22	2.88	0.98	1.51	1.16
Ethylene biosynthesis/signaling	At4g26200	253999_at	ACS7	15.45	17.22	10.19	0.95	1.18	0.89
	At3g23230	257918_at	TDR1	133.03	56.61	14.62	0.92	1.17	1.01
Jasmonic acid biosynthesis	At1g17420	261037_at	LOX3	23.74	19.48	8.25	0.93	1.14	0.95
	At1g72520	260399_at	LOX4	102.11	105.33	32.74	0.92	1.29	1.13
Indole glucosynolate biosynthesis	At1g18570	255753_at	MYB51	22.38	20.67	12.01	1.10	5.61	1.70
	At1g21100	261459_at	IGMT1	2.90	2.03	2.08	1.97	9.09	4.93
	At1g21120	261449_at	IGMT2	89.73	77.34	47.30	2.04	44.68	4.68
	At1g21110	261450_s_at	IGMT3	37.83	27.85	19.03	1.04	15.79	3.34
	At1g21130	261453_at	IGMT4	21.79	16.92	12.87	1.01	12.57	2.34
	At4g31500	253534_at	CYP83B1	3.04	1.94	2.42	1.64	2.46	1.34
Chitinase activity	At1g02360	259443_at	Chitinase family protein	2.79	4.47	2.77	1.14	2.84	1.18
	At2g43570	260568_at	Putative chitinase	0.73	1.52	1.54	4.75	21.02	4.40
	At3g54420	251895_at	EP3 class chitinase	3.95	9.75	3.28	2.36	3.29	1.42
	At4g01700	255595_at	Chitinase family protein	2.71	6.28	2.99	3.50	7.48	2.25

Not All Responses Elicited by Other PAMPs Are Elicited by Cellobiose

Cellobiose Treatment Increases Plant Growth

The global suppression of gene expression for photosynthesis genes following biotic stress is well documented, presumably as a compensatory mechanism for the high metabolic cost of defense (Bilgin et al., 2010; Göhre et al., 2012). Accordingly, at 3 h posttreatment for all three elicitors tested, photosynthesis-related genes, particularly those coding for proteins in PSI and PSII reaction centers (Table II), showed reduced expression in seedlings. Typically, exposure to high concentrations of elicitors halts seedling growth (Gómez-Gómez et al., 1999; Zipfel et al., 2006). This growth inhibition phenotype has been exploited successfully for the identification of mutants insensitive to elicitor treatments. In contrast, we observed that seedlings grown in high concentrations of cellobiose displayed increased fresh weight when compared with those grown in lower concentrations or without cellobiose (Fig. 6). It is possible that cellobiose is cleaved by β-glucosidases, either in the apoplast or in the cytoplasm, thereby increasing the cell’s availability of Glc. We are unaware of a cellobiose transporter in plants, as found in other organisms (e.g. the CDT-1 and CDT-2 transceptors in Neurospora crassa; Galazka et al., 2010). In addition, we did not observe in our microarray experiments a significant induction of the expression of genes encoding sugar transporters exclusively in cellobiose-treated samples that could suggest cellobiose/cellulose oligomer-specific transport. However, a gene encoding the β-glucosidase BGLU27 (At3g60120), a family 1 glucosidase predicted to reside in the cytoplasm (Tanz et al., 2013), was highly up-regulated exclusively in the cellobiose-treated samples (Supplemental Table S1). This result was confirmed by qRT-PCR (Fig. 6). We obtained a T-DNA insertion line of BGLU27 (Salk_005337C) in which the mRNA for this gene is reduced to undetectable levels (Fig. 6). When these lines were treated with cellobiose, we did not observe any significant changes in WRKY30 up-regulation relative to the wild type, indicating that BGLU27 is not required for cellobiose perception or signal transduction (Fig. 6). However, the plants impaired in BGLU27 expression did not grow as well in the presence of cellobiose compared with wild-type plants (Fig. 6), suggesting that BGLU27 might be a β-(1,4)-hydrolase involved in cellobiose breakdown to increase Glc availability. Importantly, in the bglu27-1 mutant background, which seems less capable of consuming cellobiose, excess cellobiose still did not have a detrimental effect on seedling growth.

Table II. Elicitor-induced fold changes of selected genes involved in photosynthesis and related metabolism 3 h after treatment.

CB, Cellobiose; CH, chitin.

Process	Gene Identifier	Probe	Gene Name/Annotation	Fold Change (P < 0.05)
				CB	CH	OG
Photosynthetic-related metabolism	At1g32470	260704_at	CDH3	−2.77	−3.56	−2.71
	At5g36700	249658_s_	PGLP1	−4.29	−5.85	−4.05
	At3g55330	251784_at	PPL1__PsbP-like protein 1	−2.40	−2.52	−2.12
PSII	At2g30570	267526_at	PSBW__photosystem II reaction center W	−2.51	−3.98	−2.41
	At3g21055	256979_at	PSBTN__photosystem II subunit T	−2.78	−3.72	−3.32
PSI	At4g28750	253738_at	PSAE-1	−3.29	−4.91	−3.34

Figure 6.

Effects of cellobiose on seedling growth. A, Fifteen-day-old Arabidopsis seedlings display increased fresh weight when grown on high concentrations of cellobiose. Plants impaired in BGLU27 expression do not grow as well in the presence of cellobiose compared with the Col-0 control. Letters indicate P < 0.05 by one-way ANOVA coupled to Tukey’s test. B, The T-DNA insertion line of BGLU27 (bglu27-1) is not impaired in cellobiose (CB)-induced (100 μm) WRKY30 expression. C, BGLU27 mRNA was not detected by qRT-PCR in bglu27-1 lines. Error bars represent sd.

Responses to Cellobiose Are BAK1 Independent

The plant Leu-rich repeat receptor kinase BAK1/SERK3 is involved in brassinosteroid hormone responses, cell death control, and innate immunity (Chinchilla et al., 2007, 2009). BAK1 has been shown to associate with Leu-rich repeat-type PRRs and to be required for signal transduction following the perception of PAMPs, including flg22 and elf18 (Chinchilla et al., 2007; Heese et al., 2007; Roux et al., 2011; Schwessinger et al., 2011). One well-studied example is BAK1 recruitment to the flagellin receptor complex following flg22 perception (Chinchilla et al., 2007; Heese et al., 2007). Plants defective in the BAK1 gene are less sensitive to flg22 treatment. We used the bak1-5 allele of BAK1, which is impaired specifically in innate immune signaling (Schwessinger et al., 2011), to assess whether BAK1 is required for signal transduction following cellobiose treatment. Plants carrying the bak1-5 mutation were impaired significantly in WRKY30 expression following flg22 treatment; however, we did not observe any changes in WRKY30 expression after cellobiose elicitation (Fig. 7). This result shows that BAK1 is not required for cellobiose perception and subsequent signal transduction.

Figure 7.

Cellobiose-induced WRKY30 expression is independent of BAK1. In contrast, WRKY30 up-regulation is reduced significantly in the bak1-5 mutant after flg22 treatment. Means and sd of three biological replicates are shown.

Responses to Cellobiose Are Independent of ROS

One of the early defense responses triggered by PAMP/DAMP recognition is the production of ROS (e.g. superoxide and hydrogen peroxide). Genetic analysis in Arabidopsis demonstrated that the ROS burst is dependent on NADPH oxidases as well as AtRbohD and AtRbohF, respiratory burst oxidase homologs of the human neutrophil gp91phox. Plants defective in RbohD and RbohF are impaired in full ROS production in response to elicitor treatment and pathogen attack (Simon-Plas et al., 2002; Torres et al., 2002, 2006; Nühse et al., 2007; Galletti et al., 2008). To determine if cellobiose treatment elicited ROS production, we used a luminol-based assay to quantify hydrogen peroxide in leaf discs treated with cellobiose. We were not able to detect any ROS signal following cellobiose treatments ranging from 100 μm up to 1 mm (Fig. 8). In addition, we transformed plants defective for both NADPH oxidase D and F (rbohD and rbohF) with the WRKY30p:GUS construct and treated homozygous T3 plants with cellobiose. We did not observe any differences in GUS expression following cellobiose treatment, indicating that AtRbohD and AtRbohF are not required for signal transduction leading to WRKY30 up-regulation in response to cellobiose (Fig. 8). A second potential source of ROS are apoplastic peroxidases. To confirm that ROS production was not necessary for the signal transduction of cellobiose perception, we took advantage of previously characterized transgenic Arabidopsis plants expressing an antisense cDNA encoding a type III peroxidase, FBP1, impaired in the oxidative burst (Bindschedler et al., 2006). Cellobiose-treated FBP1 seedlings were unaltered in WRKY30 up-regulation relative to the wild-type control (Fig. 8). Together, these experiments indicate that the cellobiose signaling pathway is independent of ROS formation.

Figure 8.

Cellobiose treatment does not elicit ROS production. A, Luminol-based assay results show no detectable ROS formation after cellobiose treatment. Means and se are shown for 20 biological replicates. B, Results of cellobiose-treated wild-type (Col-0) and rbohD/F Arabidopsis seedlings carrying the WRKY30p:GUS construct (lane 1, control; lane 2, 100 μm cellobiose) showing that cellobiose-induced WRKY30 expression in seedling roots is not impaired in the rbohD/H mutant background. C, WRKY30 relative expression measured by qRT-PCR. Arabidopsis plants expressing an antisense cDNA encoding French bean peroxidase1 (FBP1) are not impaired in cellobiose (CB) induction of WRKY30 expression.

The Expression of Genes Involved in Suberin Biosynthesis Is Induced following Cellobiose Treatment

Plant cell wall reinforcements, which occur following cell wall disruption, are typical responses following PAMP/DAMP perception and, in some cases, can be beneficial to plants. For example, callose deposition at the cell wall usually can be observed in roots and leaves in response to pathogen cell wall penetration or PAMP perception (Galletti et al., 2008; Millet et al., 2010). Overexpression of PMR4 (synonym = GSL5), encoding a stress-induced callose synthase, demonstrated that early callose deposition results in complete penetration resistance to powdery mildew in Arabidopsis (Ellinger et al., 2013). It is possible that cellobiose perception, as an indicator of cell wall damage, also leads to cell wall reinforcement. After treating plants with cellobiose for 24 h, we could not observe any callose formation (Fig. 9) or ectopic lignification (data not shown). However, when investigating our cellobiose-induced gene expression data sets, we observed elevated transcript levels for suberin biosynthetic genes at 3 h posttreatment (Fig. 10). The transcripts for cytochrome P450 CYP86B1 and the acyltransferase GPAT5 were elevated more than 3-fold relative to untreated controls. Transcript levels for other genes in this pathway, LACS2 and FAR5, were 2-fold higher. In addition, MYB41, encoding a transcription factor recently shown to activate suberin biosynthesis (Kosma et al., 2014), was induced in the cellobiose-treated samples at 25 min but not at 3 h. The suberin biosynthetic genes and MYB41 were only up-regulated by cellobiose and not by chitooligomers or OGs. In subsequent qRT-PCR experiments, we observed that all genes in the suberin pathway had peak expression 1 h after cellobiose treatment (Fig. 10), and we confirmed that these genes were not up-regulated after chitooligomer treatment (Supplemental Fig. S6). Suberin is a cell wall-linked polymer that acts as a hydrophobic barrier and is deposited in response to biotic and abiotic stresses (Thomas et al., 2007; Kosma et al., 2014). Studies have shown that the rate of tissue suberization after wounding correlates with increased resistance to subsequent fungal infection at wound sites (Biggs and Miles, 1988; Lulai and Corsini, 1998). Despite several attempts, we could not detect elevated suberin in seedling roots treated with cellobiose (data not shown). This result indicates that the up-regulation of the suberin biosynthetic pathway, triggered by cellobiose perception alone, is not sufficient to cause ectopic suberin deposition. However, it is possible that cellobiose perception may participate in preparing the plant for suberin deposition following the recognition of additional stress signals.

Figure 9.

Cellobiose (500 μm) does not induce callose accumulation in 7-d-old seedling roots. The top row shows bright-field images, and the bottom row shows UV epifluorescence images. Cell wall callose reinforcements were detected in seedlings treated with flg22 (1 μm).

Figure 10.

Expression profiles of suberin biosynthesis-related genes in seedling roots after cellobiose treatment. A, Microarray results show the up-regulation of MYB41 in the cellobiose samples at 25 min and an increase in the expression of genes in aliphatic suberin biosynthesis at 3 h after cellobiose treatment. B, Time-course expression analysis done by qRT-PCR showing the peak expression of suberin biosynthesis-related genes at 1 h after cellobiose treatment. Error bars represent sd (n = 6).

Cellobiose Pretreatment Confers Increased Resistance to Pseudomonas syringae pv tomato DC3000 Infection

Exposure to avirulent pathogens or PAMP/DAMP elicitor treatment can prepare the plant’s immune system for a more efficient defense reaction to subsequent pathogen attacks (Van Wees et al., 2008). Given the short-lived PTI signaling observed after cellobiose treatment and the apparent lack of ROS formation, we were interested in investigating whether cellobiose treatment could induce increased resistance against pathogen attack in Arabidopsis plants. We compared ion leakage, used as a proxy for P. syringae pv tomato DC3000-induced cell leakage and death, in infected plants pretreated with water, cellobiose, or flg22. The results showed that plants pretreated with cellobiose were more resistant to infection than plants pretreated with water, and the resistance effect was not significantly different from that provided by flg22 pretreatment (Fig. 11). However, pretreatment with flg22 seemed to confer a stronger protection against P. syringae pv tomato DC3000, since ion leakage in those plants was not significantly different from the ion leakage measured in uninfected controls (Fig. 11).

Figure 11.

Analysis of ion leakage in 2-week-old Arabidopsis seedlings after infection with P. syringae pv tomato DC3000 via flood inoculation. The y axis shows ion leakage relative to the total ion content, and the x axis shows pretreatments: deionized water, 500 μm cellobiose (CB), or 10 μm flg22. The uninfected control was pretreated with water. Letters indicate P < 0.05 by one-way ANOVA coupled to Tukey’s test. Error bars represent sd (n = 12).

Cellobiose Has an Additive Effect with Other PAMPs/DAMPs on PTI Signaling

About 2.4% of the Arabidopsis genes encode receptor-like kinases, some of which function as PRRs at the cell surface. This large number of receptors may reflect the variety of eliciting signals that plants can perceive (Boller and Felix, 2009; Macho and Zipfel, 2014). We were interested in investigating the independent nature of cellobiose perception and if the signal cascade generated by cellobiose perception traveled through similar pathways as for other known PAMPs/DAMPs. We investigated PTI signaling outputs in combination treatments, in which cellobiose was applied simultaneously with another elicitor (i.e. flg22, chitooligomers, or OGs). We were able demonstrate that the calcium spike generated by cellobiose is independent of and/or additive to that of other elicitors (Fig. 12). In addition, by comparing the calcium signatures derived from the different elicitor treatments, we observed that the calcium signature generated by cellobiose is similar to that of pectin-derived OGs, as opposed to the slightly delayed and longer lasting curve generated by treatment with the PAMPs flg22 and chitooligomers. In particular, the calcium signature from the simultaneous application of cellobiose and flg22 was a curve distinct from and higher in amplitude than the calcium signatures of either cellobiose of flg22 single treatments, indicating perhaps that each elicitor has a different mode of triggering changes in intracellular calcium levels.

Figure 12.

Combination treatments of cellobiose (CB) together with other elicitors. A, Intracellular calcium influx was measured in aequorin-expressing seedlings and immediately visualized using a CCD camera. Means and sd are shown. At least nine biological replicates were measured per treatment. The pixel intensities of images were captured and then quantified using ImageJ. B, Cellobiose treatment in combination with chitooligomers or flg22 increased the intensity and duration of MPK activation profiles relative to individual treatments. C, Amplification of WRKY30 expression in seedling roots also was observed in combination treatments. Treated samples were compared with untreated controls grown in parallel. Letters indicate P < 0.05 by one-way ANOVA coupled to Tukey’s test. Error bars represent sd from three biological replicates. The experiment was repeated twice with similar results.

MAPK activation was amplified in combination treatments of cellobiose, chitooligomers, and flg22. In single-treatment experiments with chitooligomers and flg22, peak activation of MAPK was obtained at 30 min posttreatment, with little activation visible at 60 min. In treatments of cellobiose combined with flg22 or chitooligomers, MAPK activation was stronger at 30 min and still visible at 60 min (Fig. 12). We also observed an amplification of expression of the marker gene WRKY30 in samples treated with cellobiose combined with either chitooligomers or flg22 after 25 min (Fig. 12). Current evidence, including the work presented here, suggests that the initial phase of danger signaling triggers a response similar in qualitative terms although quantitatively different according to the particular danger signals involved (Denoux et al., 2008; Boller and Felix, 2009). Our results with combination PAMP/DAMP treatments suggest an independent mode of cellobiose perception and clearly show a quantitative amplification of the immune signaling cascade. The amplification of defense signaling in response to the simultaneous perception of multiple stimuli may render a stronger immune response.

DISCUSSION

Plant cell walls are a source of potential defense signaling molecules that can be released upon degradation by pathogen enzymatic repertoires (Hahn et al., 1981; Walton, 1994). Upon the perception of cell wall damage, cells respond by activating signaling cascades leading to the activation of defense responses. We used WRKY transcription factors as defense markers to identify cell wall oligosaccharides capable of activating defense responses in Arabidopsis. We showed that Arabidopsis can perceive cellulose degradation products like cellobiose and respond by activation of a signaling cascade leading to the increased expression of defense-related genes, with substantial overlap relative to other pathogen-associated and cell wall damage-associated elicitors. Cellobiose pretreatment induced in Arabidopsis seedlings an immune response that resulted in less cell damage following P. syringae infection. Cellobiose treatment caused a rapid and transient intracellular calcium spike, which was similar to the timing and shape of the calcium response to OGs. When treating aequorin-expressing seedlings with a combination of elicitors, we observed an additive or synergistic effect in the calcium signatures, most noticeably for cellobiose plus flg22. Despite the critical nature of Ca²⁺ signaling to pathogen defense, there is still a limited mechanistic understanding of how different calcium signatures affect gene expression and defense outcomes (Seybold et al., 2014). While some studies suggest an apoplastic origin of PAMP-induced Ca²⁺ influx (Aslam et al., 2008; Ranf et al., 2011; Segonzac et al., 2011), other researchers propose a requirement for intracellular Ca²⁺ stores (Ma et al., 2012). It is possible that concurrent PAMP/DAMP perception may lead to synergistic changes in Ca²⁺ signaling signatures, resulting in increased immune fitness. Treatment with cellobiose activates MAPKs at very early time points and appears to be ROS independent. Recent findings using chemical genetic approaches showed that the oxidative burst and MAPK activation are two independent signaling events in plant immunity (Ranf et al., 2011; Segonzac et al., 2011; Xu et al., 2014), which is in agreement with our results. The transient nature of the responses triggered by cellobiose suggests that the perception of cellobiose may be auxiliary to other stimuli. During a pathogen attack, PAMP perception, the detection of cell wall breakdown, membrane distortion, and depolarization all may contribute to the intensity of plant responses. Current research on PTI focuses on responses to single elicitors, an unlikely scenario in nature. Our results from treatments with two elicitors show that a number of signaling steps in PTI are enhanced, suggesting that plants may be able to respond to lower elicitor levels and more quickly with effective defenses than previous work has indicated.

Cellulose microfibrils help provide the tensile strength that dictates the structure of the plant cell (Somerville et al., 2004). Drastic loss of cellulose microfibril integrity leading to changes in cell shape and size elicits defense-like changes in gene expression. Mutants defective in cellulose synthesis, such as the CESA3 mutants cev1 and eli1-2, exhibit increased resistance to powdery mildew pathogens due to increased activation of defense hormone signaling, induction of defense response gene expression, and increased cell wall reinforcement by lignification (Ellis and Turner, 2001; Caño-Delgado et al., 2003). In addition, the transmembrane malectin receptor kinase THESEUS has been shown to mediate the ability of plants to respond to defects in cellulose disruption observed in cesA6 mutants but does not participate in cellobiose perception (Hématy et al., 2007; C. de Azevedo Souza and S.C. Somerville, unpublished data). Together, these studies highlight the role of the cellulosic fraction of the plant cell wall in generating signals activating a cell wall integrity system. Cellobiose fragments are likely generated by cellulase digestion of cellulose but prior to the collapse of cell wall integrity. Thus, we could not detect any clear evidence of cell wall reinforcement following overnight cellobiose treatment; however, we did observe the up-regulation of genes required for suberin biosynthesis exclusively in the cellobiose-treated samples, suggesting a possible link between cellobiose perception and cell wall reinforcement through suberin deposition. Our data suggest that plants can directly monitor the status of cellulose by perceiving small oligomers of cellulose. It is possible that this perception is mediated by PRRs, similar to other PAMPs/DAMPs, but the identity of the putative receptor and detailed molecular mechanisms of perception and signal transduction are unknown. The rapid calcium influx and MAPK activation observed suggest a receptor-mediated perception at the cell surface. The relatively high levels of cellobiose (greater than 100 μm) required to obtain detectable readouts, 100-fold more than what is required for flg22-triggered responses, suggest that a putative membrane receptor dedicated to cellulose oligomer perception must have low sensitivity, perhaps to account for the cellulose fragments that may be generated during cell wall remodeling, thus preventing unnecessary stress responses. However, it is also possible that cellobiose may be internalized or perceived indirectly, for example, serving as a donor molecule for the modification of other molecules prior to the activation of defense responses.

This work demonstrates that Arabidopsis can perceive breakdown products of the cellulosic fraction of the plant cell wall, and this perception, concurrent with the perception of PAMPs, enhances downstream defense signals. We are currently working to identify the molecular components involved in cellobiose perception.

MATERIALS AND METHODS

Plant Material and Growth Conditions

Arabidopsis (Arabidopsis thaliana) accession Col-0 was the background for all mutants and transgenic lines used in this study. Seeds were surface decontaminated with a 30% bleach solution in 0.1% SDS with agitation for 15 min. Seeds were subsequently washed three times with distilled water and then stratified for at least 3 d at 4°C. Individual seeds were placed in separate wells of a flat-bottom transparent 96-well plate covered with plastic wrap and grown on Murashige and Skoog (MS; Caisson Laboratories) liquid medium (1× MS salts, 2.5 mm MES, and 0.5% [w/v] Suc, pH 5.7). For RNA extraction experiments, approximately 50 to 100 seeds were sown on a 125-μm-aperture nylon mesh (Industrial Netting) and floated over liquid MS medium. For MAPK experiments, 15 seeds were added to wells of a 12-well plate. Plants were grown in growth chambers (Percival CU36L5) with 24 h of light of 120 µmol m⁻² s⁻¹ (400- to 700-nm range) provided by fluorescent F17T8/TL741 (ELA-039) bulbs and at a constant temperature of 22°C.

Elicitors

Chitin oligomers from hydrolyzed shrimp shells was obtained from Sigma-Aldrich (catalog no. C9752), oligogalacturonans (DP12–DP25) were obtained from F.M. Ausubel, and the flg22 peptide (amino acid sequence QRLSTGSRINSAKDDAAGLQIA) was synthesized by Elim Biopharmaceuticals at a purity level of 70% or greater. Stocks of flg22 were prepared by dissolving the peptide in water at a concentration of 10 mm and stored at −20°C. Chitooligomers were dissolved in water at 10 mg mL⁻¹, autoclaved, and centrifuged to remove insoluble materials. Cellobiose was obtained from Fluka (catalog no. 22150). All other sugars used in this study were obtained from Megazyme. Seven-day-old seedlings were treated with PAMP/DAMP elicitors added to MS liquid medium. Unless noted otherwise, typical treatments consisted of eliciting molecules at saturating concentrations: 100 μg mL⁻¹ chitin, 100 μg mL⁻¹ OGs, 1 μm flg22 (Felix et al., 1999; Hu et al., 2004; Miya et al., 2007; Shinya et al., 2012), and 100 μm cellobiose.

Generation and Analysis of GUS Reporter Lines

GUS reporter lines of WRKY transcription factors were created using Gateway technology (Invitrogen). Promoter sequences of about 2 kb in length were PCR amplified from Arabidopsis Col-0 genomic DNA and cloned into vector PGWB3 upstream of the GUS open reading frame. The resulting plasmids were transferred into Col-0 plants by Agrobacterium tumefaciens (GV3101)-mediated transformation (Clough and Bent, 1998). Homozygous transformants were grown in liquid medium and inspected for GUS expression after various treatments as indicated. Seven-day-old seedlings were treated with elicitors for 16 h, placed in the GUS substrate solution (50 mm sodium phosphate buffer, pH 7, 0.1% Triton X-100, 3 mm potassium ferricyanide, 3 mm potassium ferrocyanide, and 1 mm 5-bromo-4-chloro-3-indolyl β-d-glucuronide), and incubated for 8 to 16 h at 37°C (Jefferson, 1987). Seedlings were mounted on glass slides with 25% glycerol and imaged using a photoscanner (Epson Perfection V600 Photo).

Isolation of Seedling Root Tissue

Seedlings were grown in liquid MS medium suspended over a nylon mesh as noted above, in which roots passed through the mesh apertures (125 μm nominal hole size), allowing for the separation of roots and shoots. Mesh discs containing 7-d-old seedlings (Supplemental Fig. S2) were transferred to petri plates for the various treatments and frozen immediately thereafter. Seedling roots were broken off the mesh and homogenized for RNA extraction.

Reverse Transcription-PCR and qRT-PCR

Total RNA was extracted from homogenized tissue frozen in liquid nitrogen and digested with DNase (catalog no. 79254; Qiagen), and 1 μg of RNA per 20-μL reaction was used to generate first-strand cDNA using SuperScript II Reverse Transcriptase (Invitrogen) following the manufacturer’s protocol. For reverse transcription-PCR analysis of WRKY30 expression in seedling roots, gene-specific and intron-spanning primers (Supplemental Table S3) were used in PCR to amplify corresponding cDNA sequences under the following PCR conditions: 95°C for 3 min, 28 cycles of 94°C for 30 s, 57°C for 30 s, and 72°C for 1 min, and 72°C for 4 min, using Taq polymerase (Clontech Laboratories) in a 25-μL reaction. PCR products were separated on 1% ethidium bromide agarose gels and photographed under a UV transilluminator (Bio-Rad Gel Doc XR). ACTIN1 was used as a control (Supplemental Table S3). For qRT-PCR experiments, cDNA was obtained as described above, and 1 μL was used to analyze gene expression using SYBR GreenER qPCR Supermix (Life Technologies) and the following PCR conditions: 50°C for 2 min, 95°C for 10 min, 40 cycles of 95°C for 15 s, 59°C for 30 s, and 68°C for 45 s, followed by a fluorescence reading. The housekeeping control rRNA 60S (Walley et al., 2007) was amplified in parallel on each plate for normalization. No-template controls and melting curves were examined to ensure against contamination and primer-dimer formation. The relative starting quantities of each gene were determined by the ΔΔCT method (Hietala et al., 2003). Unless noted otherwise, primers were designed using the online tool ATRTPrimers (Han and Kim, 2006), and primers spanning exon-intron boundaries were selected whenever possible. Primers are listed in Supplemental Table S3.

Calcium Measurements

Relative intracellular calcium influxes after elicitor treatment were measured using an aequorin-based calcium assay (Knight et al., 1996; Tanaka et al., 2010). In short, individual 6-d-old aequorin-expressing seedlings were transferred to individual wells of a 96-well microplate and incubated overnight in reconstitution buffer containing coelenterazine (catalog no. 55779; Biosynth International). Since the timing of the response is critical, solution trays with three wells were used to separate individual treatments and allow concurrent dispensing using a multichannel pipettor. Plants were measured immediately in a luminescent image analyzer (LAS4000; Fuji Film), using a 50-s integration time with 10 repetitions, for a total of 500 s per sample. Nine to 12 biological replicates were used for each treatment, and each set of treatments was repeated at least three times. Images were analyzed using ImageJ (http://imagej.nih.gov/ij) for the measurement of pixel intensity.

MAPK Assays

MAPK assays were performed as described previously with minor modifications (Tsuda et al., 2009). Arabidopsis seedlings were grown for 7 d on 12-well plates (15 seedlings per well) in which each well contained 3 mL of liquid MS medium with 0.5% Suc. Elicitors were added, and seedlings were harvested at different time points as indicated and immediately frozen in liquid nitrogen. The frozen seedlings were ground in liquid nitrogen and homogenized in 100 µL of extraction buffer: 100 mm HEPES, pH 7.5, 5 mm EDTA, 5 mm EGTA, 2 mm DTT, 10 mm Na₃VO₄, 10 mm NaF, 50 mm β-glycerol phosphate (Santa Cruz Biotechnology), 1× proteinase/phosphatase inhibitor cocktail (Cell Signaling Technology), 10% glycerol, and 1% (w/v) polyvinylpolypyrrolidone. After centrifugation at 16,000g for 30 min at 4°C, supernatants were frozen and stored at −20°C. The protein concentration was determined using a Bradford assay (Bio-Rad) with BSA as a standard. Protein (20 µg) was separated on a 12% polyacrylamide gel. Immunoblot analysis was performed using anti-phospho-p44/42 MAPK (1:2,000; Cell Signaling Technology) and anti-AtMPK3 (1:2,000; Sigma-Aldrich) as primary antibodies and peroxidase-conjugated goat anti-rabbit IgG (1:15,000; catalog no. A 6154; Sigma-Aldrich) as a secondary antibody.

Microarray Experiments

RNA was extracted from roots of seedlings treated with cellobiose, chitooligomers, OGs, or a no-elicitor control, for 25 min and 3 h, using Trizol LS (Invitrogen) according to the manufacturer’s recommendations. RNA integrity was checked with an Agilent 2100 BioAnalyzer. An aliquot from each RNA sample was used as a template to make cDNA, which was assessed by qRT-PCR to confirm that samples had the expected WRKY30 expression profile at 25 min. Samples were then analyzed for gene expression with Affymetrix GeneChip Arabidopsis ATH1 Genome Arrays using standard Affymetrix reagents and protocols at the QB3-Functional Genomics Laboratory at the University of California, Berkeley. Samples from three biological replicates for each treatment were analyzed. A total of 24 chips were used (4 treatments × 2 time points × 3 biological replicates). Microarray data were analyzed using the GCRMA algorithm as described previously (Fletcher et al., 2011); ratios of normalized probe set intensity values were calculated for each sample pair (in which the M value = log₂ [elicitor/control]) and then averaged among the three replicates. Average linkage hierarchical analysis of arrays was performed using Cluster 3.0 and visualized using Java Treeview (Saldanha, 2004). Venn diagrams of differentially expressed genes with fold change of 2.5 or greater were generated using the MapMan version 10.0 software package (Thimm et al., 2004).

ROS Measurements

Oxidative burst measurement was performed using a luminol-based assay (Gimenez-Ibanez et al., 2009). ROS was elicited with chitooligomers or cellobiose, and elicitation in the absence of any PAMP (water treatment) was included in all experiments as a negative control. Twenty leaf discs from 10 5-week-old Col-0 plants were used for each condition. Luminescence was measured over time using an ICCD photon-counting camera (Photek).

Infection Assays

Fifteen-day-old Arabidopsis Col-0 seedlings were tested for resistance against Pseudomonas syringae pv tomato DC3000 (Whalen et al., 1991). Disease-associated water soaking was estimated by measuring ion leakage 3 d postinoculation (Potnis et al., 2015; Ishiga et al., 2016). Infections were performed via the flood inoculation method (Ishiga et al., 2011) with minor modifications. Plants were grown in MS medium solidified with 0.5% Phytagel (Sigma-Aldrich) in 237-mL sterile culture vessels (PhytoTechnology Laboratories). Plants were pretreated twice by flooding for 3 min with 50 mL of water, 500 μm cellobiose, or 10 μm flg22 at 24 and 4 h prior to infection. Plants were infected by flooding the chamber with 50 mL of 1 × 10⁵ colony-forming units of bacterial suspension in sterile water containing 0.025% Silwet. All treatments were at room temperature. Aerial parts of inoculated seedlings and uninfected controls were harvested at 3 d after infection. Four rosettes were harvested per treatment, placed individually in culture tubes filled with 6 mL of distilled water, and gently agitated for 3 h. Plants were then transferred into a new tube containing 6 mL of distilled water and autoclaved for 30 min to release total ions. Leachates were measured using an ion conductivity meter (Thermo Orion model 105, with conductivity cell 011050). Values relative to the whole ion content were used to express percentage ion leakage. Each experiment consisted of four replications, and the experiment was performed three times independently.

Callose Staining

Arabidopsis Col-0 seedlings were grown on plates containing one-half-strength MS medium supplemented with 1% agar and grown vertically in growth chambers as described above. Seven-day-old seedlings were treated with 100 μm cellobiose and 1 μm flg22 overnight. Elicitor-treated seedlings were incubated in Aniline Blue staining solution (0.01% Aniline Blue in 150 mm K₂HPO₄, pH 9.5) for 4 h (Adam and Somerville, 1996), subsequently mounted on microscope slides in 25% glycerol, and then observed with a Leica DMI 5000 B epifluorescence microscope with a 20× objective and A4 filter set (365- ± 25-nm excitation filter, 400-nm dichroic, and 450-nm long-pass emission filter).

Accession Numbers

Microarray data were deposited in the NCBI Gene Expression Omnibus under accession number GSE87217 (Edgar et al., 2002). Seeds of lines pW30:GUS (accession no. CS69613), pW40:GUS (accession no. CS69614), and homozygous bglu27-1 Salk_005337C line (accession no. CS69615) were deposited in the Arabidopsis Biological Stock Center (Alonso et al., 2003).

Supplemental Data

The following supplemental materials are available.

• Supplemental Figure S1. Induction of WRKY30 expression with oligomers of cellulose of DP2 to DP4.

• Supplemental Figure S2. WRKY30 expression in seedling roots versus shoots after PAMP/DAMP treatments.

• Supplemental Figure S3. Effect of cerk1 mutation on WRKY30 expression.

• Supplemental Figure S4. Calcium influx control experiments.

• Supplemental Figure S5. MPK identification experiment.

• Supplemental Figure S6. Expression profile of suberin biosynthesis-related genes after chitooligomer treatment.

• Supplemental Table S1. Genes up-regulated over 2.5-fold 25 min after PAMP/DAMP treatment.

• Supplemental Table S2. Genes up-regulated over 2.5-fold 3 h after PAMP/DAMP treatment.

• Supplemental Table S3. Primers used in this study.

Glossary

PAMP

pathogen-associated molecular pattern

PRRs

pattern recognition receptors

PTI

pattern-triggered immunity

ROS

reactive oxygen species

DAMP

damage-associated molecular pattern

OG

oligogalacturonide

qRT

quantitative reverse transcription

Col-0

Columbia-0

MS

Murashige and Skoog